Writing in Nature, Lizzie Buchen outlines how the idea that the parents of autistic children, and the children themselves, have an aptitude for understanding and analysing predictable, rule-based systems came to be; the genes that endow parents with minds suited to technical tasks could lead to autism when passed on to their children, especially when combined with a dose of similar genes from a like-minded mate. And how Simon Baron-Cohen promoted the idea.
It has an air of truthiness, as Simon Baron-Cohen always does, and it is probably a realistic premise if you have no clue about biology.
What famous scientist isn't (or wasn't) autistic when the criteria become so broad? I can't think of any. The downside is that while it gets headlines for Baron-Cohen, it does a real disservice to parents who are dealing with children who have real autism but who will be increasingly regarded as engaging in faddish pop psychology diagnoses.
The problem is the same thing we see in too many psychology studies that draw conclusions; they use self-reported surveys. So people can object to concerns that psychology in the last few decades lacks the rigor to be called science, but it would sure help if they would continue to hold famous people in psychology accountable to using scientific rigor.
Scientists and autism: When geeks meet - by Lizzie Buchen published online 2 November 2011, Nature 479, 25-27 doi:10.1038/479025a